Predictive Simulations to Replicate Human Gait Adaptations and Energetics With Exoskeletons

Area of Science:

• Biomechanics
• Robotics
• Human-Computer Interaction

Background:

• Robotic exoskeletons offer potential for mobility restoration and enhancement.
• Optimal control of exoskeletons in conjunction with human users remains a significant challenge.
• Accurate simulations of human-exoskeleton interaction can accelerate design and improve control strategies.

Purpose of the Study:

• To validate predictive simulations for replicating human gait adaptations and energy expenditure changes during exoskeleton use.
• To demonstrate the utility of simulations in understanding human responses to altered energetic landscapes imposed by exoskeletons.

Main Methods:

• Recreated a prior experimental setup using knee-worn robotic exoskeletons applying frequency-dependent resistive torques.
• Developed predictive models to simulate exoskeleton-human interaction and energy expenditure.
• Varied virtual participant anthropometrics and muscle parameter sets to assess prediction accuracy.

Main Results:

• Simulations successfully replicated experimental findings of altered optimal step frequencies (higher and lower than natural) under different exoskeleton conditions.
• Modeled resistive torques and energy landscapes closely matched experimental data.
• Individual muscle energetics indicated distinct coordination strategies corresponding to exoskeleton controller types.
• Increasing the number of virtual participants improved prediction accuracy more than increasing muscle parameter sets.

Conclusions:

• Predictive simulations can accurately replicate human gait adaptations and energy changes when using robotic exoskeletons.
• The simulation approach aids in understanding user coordination strategies and optimizing exoskeleton controller design.
• This method facilitates pre-human testing controller design, identifying solution spaces, and tailoring designs for individual users.